Knowledge graph construction and analysis system for intelligent electric meter data

Abstract

The invention discloses a knowledge graph construction and analysis system for intelligent electric meter data. The knowledge graph construction and analysis system comprises the following units: a knowledge graph construction unit, an anomaly detection unit, a graph reasoning unit, a diagnosis report generation unit and an intelligent alarm unit, multi-dimensional electric power data of the intelligent electric meter are obtained and preprocessed through a knowledge graph construction unit, and then an electric power field knowledge graph associated with electric meter equipment information, electric power parameter data, fault records and equipment association relationships is constructed based on the preprocessed data. According to the knowledge graph construction and analysis system for the data of the intelligent electric meter, by constructing the knowledge graph in the electric power field, the system can perform comprehensive preprocessing and correlation analysis on multi-dimensional electric power data of the intelligent electric meter and accurately capture the relationship among electric power parameters, equipment information and fault records; and the accuracy of anomaly detection and fault reasoning is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electric power data processing and analysis, more particularly to a knowledge graph construction and analysis system for smart meter data. BACKGROUND

[0002] With the popularization of smart meters in the power system, real-time collection and analysis of electric power data has become an important means to improve the efficiency of power grid operation and ensure the safety of power supply. Smart meters can record electric power data in real time, including current, voltage, power and other parameters, and upload them to the power company for analysis through the communication network. However, with the sharp increase in the amount of electric power data, traditional data analysis methods face many challenges, such as insufficient accuracy of data anomaly detection, complexity of fault reason inference, and untimely alarm processing.

[0003] Existing smart meter data analysis systems mostly focus on single data processing or anomaly detection algorithms, lack of deep mining and fusion of complex relationships between electric power data, leading to false positives or false negatives in large-scale electric power data monitoring, and difficulty in diagnosing fault causes in a timely manner. Especially in dealing with complex faults and abnormalities in the power system, existing technologies have not provided enough intelligent solutions to meet the needs of efficient and accurate fault diagnosis and processing of smart grids.

[0004] Therefore, based on knowledge graph technology, a system that can comprehensively process multi-dimensional smart meter data, automatically detect anomalies, infer fault causes and generate diagnostic reports has become an urgent need to improve the level of smart meter data analysis and fault management. SUMMARY

[0005] The present application aims to provide a knowledge graph construction and analysis system for smart meter data, which solves the problem that existing smart meter data analysis systems mostly focus on single data processing or anomaly detection algorithms, lack of deep mining and fusion of complex relationships between electric power data, leading to false positives or false negatives in large-scale electric power data monitoring, and difficulty in diagnosing fault causes in a timely manner, and cannot meet the use requirements.

[0006] The present application achieves the above-mentioned purpose through the following technical solutions: a knowledge graph construction and analysis system for smart meter data, comprising the following units: a knowledge graph construction unit, an anomaly detection unit, a graph reasoning unit, a diagnostic report generation unit and an intelligent alarm unit; The multi-dimensional electric power data of the smart meter is obtained and preprocessed by the knowledge graph construction unit, and then the electric power domain knowledge graph associated with the electric meter equipment information, electric power parameter data, fault records and device association is constructed based on the preprocessed data; The anomaly detection unit fuses machine learning algorithms and knowledge graph feature to monitor power data abnormal fluctuations; The graph reasoning unit combines knowledge graph association relationships, reasoning mechanisms, and historical data to infer possible fault causes corresponding to anomalies; The diagnosis report generation unit automatically generates a diagnosis report containing anomaly details, fault causes, and treatment suggestions; The intelligent alarm unit sends corresponding intelligent alarm signals according to the anomaly level, realizing power anomaly detection and fault management.

[0007] Further, the multi-dimensional power data obtained by the knowledge graph construction unit includes: At least one of current data, voltage data, power data, power consumption duration data, and electric meter device identification information, and the data type is time series structured data.

[0008] Further, the preprocessing of multi-dimensional power data by the knowledge graph construction unit includes: Data cleaning and standardization processing; Data cleaning is used to eliminate missing values, abnormal isolated points, and invalid data caused by data collection errors; Standardization processing is used to unify the data dimensions of different power parameters.

[0009] Further, the power domain knowledge graph contains core entity types and inter-entity association relationships; The core entity types include at least one of smart meter device entities, power parameter entities, fault type entities, device association entities, and time entities; The inter-entity association relationships include at least one of collection relationships, connection relationships, trigger relationships, and corresponding relationships.

[0010] Further, the knowledge graph construction unit stores knowledge graph data using a graph database; Map the preprocessed power data to the attribute values of the corresponding entities in the knowledge graph; At the same time, calculate the association weight between entities to perfect the knowledge graph.

[0011] Further, the machine learning algorithm fused by the anomaly detection unit includes the following steps: Based on the model of integrated learning and time series anomaly detection model; The anomaly detection unit extracts entity association features, power parameter time series features, and historical fault association features from the knowledge graph to form a multi-dimensional feature set; Input the anomaly detection model constructed by the machine learning algorithm; Calculate the anomaly score and compare it with the preset threshold; Determine whether there is a power data abnormal fluctuation.

[0012] Further, when the atlas reasoning unit reasons the possible fault cause corresponding to the anomaly, the following steps are included: A reasoning rule library containing mapping rules of power parameter anomalies and fault types, device association relationships and fault conduction rules, and historical fault recurrence association rules is constructed; At least one of a path reasoning, a semantic reasoning and a probabilistic reasoning mechanism is adopted; The fault cause is reasoned in combination with historical fault handling data and real-time anomaly characteristics.

[0013] Further, the diagnostic report generation unit generates a diagnostic report, including the following steps: Key information in the anomaly detection result is extracted, including anomaly occurrence time, anomaly power parameter type, anomaly value range and anomaly confidence; The fault cause list output by the atlas reasoning unit is associated; The corresponding processing suggestion in the fault handling scheme library in the knowledge graph is matched; The information is integrated according to a preset report template to generate a structured diagnostic report.

[0014] Further, the intelligent alarm unit determines the anomaly level, including the following steps: An anomaly level score is calculated based on the anomaly confidence in the anomaly detection result and the severity coefficient corresponding to the fault cause; The anomaly is divided into at least one of a slight anomaly, a general anomaly and a serious anomaly according to the score; Different alarm modes are configured for different anomaly levels.

[0015] Further, the alarm mode includes: At least one of a system pop-up prompt, an SMS notification, an email reminder and an audible and visual alarm; The serious anomaly level triggers multi-channel synchronous alarm, the alarm signal contains a diagnostic report access portal and an anomaly core information abstract, and supports rapid positioning of the anomaly and review of the diagnostic details.

[0016] The present application has the following advantages: 1. By constructing the knowledge graph in the power field, the system can comprehensively preprocess and associate analyze the multi-dimensional power data of the smart meter, accurately capture the relationship between the power parameters, device information and fault records, and improve the accuracy of anomaly detection and fault reasoning.

[0017] 2、Combining machine learning algorithms with knowledge graphs of multi-dimensional features, the system can automatically detect abnormal fluctuations in power data, and infer possible fault causes through graph reasoning mechanisms. This intelligent detection and reasoning method can effectively reduce human intervention, improve the accuracy and response speed of fault diagnosis.

[0018] 3、According to the detected abnormal data and inferred fault causes, a structured diagnostic report is automatically generated. The report content includes abnormal details, fault causes and treatment suggestions, greatly simplifying the work of power operation and maintenance personnel and improving work efficiency.

[0019] 4、The system can automatically send alarm signals according to the abnormal level through the intelligent alarm unit, ensuring that the power equipment can notify the relevant personnel in time when the abnormality occurs. According to the severity of the abnormality, the system provides multiple alarm methods, including pop-up prompts, SMS notifications, email reminders, and sound and light alarms, to ensure the timeliness of fault handling.

[0020] 5、Through the intelligent implementation of abnormal detection and fault management, the system can effectively reduce the failure rate of power equipment and improve the safety and stability of the power system through timely abnormal early warning and fault reasoning.

[0021] 6、Based on historical fault data and real-time monitoring data, the system continuously optimizes reasoning rules and abnormal detection models, has self-learning and adaptive ability, and can cope with the changing operating environment and complex fault conditions in the power system. BRIEF DESCRIPTION OF DRAWINGS

[0022] The drawings described herein are used to provide further understanding of the present application, and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application, and do not constitute an improper limitation on the present application. In the drawings: Fig. 1 is a system block diagram of the present application; Fig. 2 is a knowledge graph construction flowchart of the present application; Fig. 3 is an abnormality detection and alarm flowchart of the present application. DETAILED DESCRIPTION

[0023] It is necessary to point out here that the following specific embodiments are only used to further illustrate the present application, and cannot be understood as a limitation on the scope of protection of the present application. Those skilled in the art can make some non-essential improvements and adjustments to the present application based on the above application content. EMBODIMENT

[0024] Please refer to Figs. 1-3The application provides a technical solution: an intelligent electric meter data knowledge graph construction and analysis system, characterized in that the system comprises a knowledge graph construction unit, an anomaly detection unit, a graph reasoning unit, a diagnosis report generation unit and an intelligent alarm unit. The system realizes power anomaly detection and fault management through the fusion of knowledge graphs and machine learning algorithms. The specific process comprises: acquiring multi-dimensional power data of an intelligent electric meter and performing preprocessing; The multi-dimensional power data of the intelligent electric meter refers to power-related information collected by the intelligent electric meter, such as voltage, current, power, power factor and other data of different dimensions. The preprocessing is to clean, convert, normalize and perform other operations on the acquired original power data to eliminate noise, missing values, outliers and other problems in the data, so that the data meets the requirements of subsequent analysis; Based on the preprocessed power data, a power field knowledge graph is constructed, and electric meter device information, power parameter data, historical fault records and device association relationships are associated; The power field knowledge graph is a knowledge graph constructed based on related concepts, entities and relationships in the power field, and is used to represent electric meter device information, power parameter data, historical fault records and device association relationships. The electric meter device information is various information about the intelligent electric meter itself, such as meter model, manufacturer, installation location, running time, etc. The power parameter data is various power parameters measured by the intelligent electric meter, such as the above-mentioned voltage, current and other specific numerical data. The historical fault record is a detailed record of the faults of the intelligent electric meter and related power equipment in the past period of time, including fault occurrence time, fault phenomenon, fault handling method, etc. The device association relationship is various associations between the intelligent electric meter and other power equipment, such as the connection relationship between a certain electric meter and a specific transformer, line, etc. The anomaly detection unit fuses machine learning algorithms and knowledge graph features to monitor abnormal fluctuations in power data in real time; The machine learning algorithm is a class of algorithm that automatically learns patterns and rules from data to make predictions and classifications on new data. In anomaly detection, these algorithms can be used to identify abnormal patterns in power data. The knowledge graph features are various entity attributes, relationships and other information features contained in the knowledge graph. These features can provide additional context information for machine learning algorithms to assist in anomaly detection. Abnormal fluctuations are changes in power data that deviate from the normal range or expected patterns, which may indicate power anomalies. The graph reasoning unit uses the association relationships and reasoning mechanisms of the knowledge graph to infer possible fault causes corresponding to the anomalies in combination with historical data; The correlation relationship is various connections between entities in the knowledge graph, such as the causal relationship between the electric meter and the fault, the connection relationship between the devices, etc.; the reasoning mechanism is based on the structure and rules of the knowledge graph, and the process of deducing unknown information from known information through logical reasoning, similarity calculation, etc.; the historical data are power-related data recorded in the past, including power parameter data, fault records, etc., used to assist in reasoning the possible fault causes corresponding to the abnormality; A diagnostic report containing abnormality details, fault causes and treatment suggestions is automatically generated by the diagnostic report generation unit; The abnormality details are a detailed description of the detected power abnormality, including the time, place, involved electric meter equipment, and specific manifestations of the abnormality; the fault cause is the possible cause of the power abnormality deduced by graph reasoning; the treatment suggestion is a corresponding solution or suggestion given for the detected abnormality and the deduced fault cause, such as checking a specific device, adjusting power parameters, etc.; The intelligent alarm unit sends corresponding intelligent alarm signals according to the abnormality level; The abnormality level is a different level divided according to the severity of the power abnormality, the range of influence, etc.; the intelligent alarm signal is a signal with specific meaning sent according to different abnormality levels, which can be in the form of sound signal, light signal, SMS notification, system pop-up window, etc., used to remind relevant personnel to handle the abnormality in a timely manner.

[0025] It should be noted that, in use, the knowledge graph construction unit can integrate multi-dimensional data, associate electric meter equipment, parameter, fault record, etc. information, form a comprehensive and structured knowledge system, provide a solid foundation for subsequent analysis, the abnormality detection unit combines machine learning algorithm and knowledge graph features, can accurately monitor abnormal fluctuations in real time, improve the accuracy and timeliness of detection, the graph reasoning unit can deeply mine the fault causes behind the abnormality with the help of correlation relationship and reasoning mechanism, combined with historical data, enhance the fault analysis capability, the diagnostic report generation unit automatically generates a report containing key information, which facilitates the staff to quickly understand the situation, the intelligent alarm unit sends corresponding signals according to the abnormality level, which can remind relevant personnel to handle in a timely manner, improve the fault response speed, the overall design realizes the intelligent and accurate power abnormality detection and fault management, which helps to ensure the stable operation of the power system.

[0026] In an embodiment, multi-dimensional power data of the smart meter are acquired and preprocessed, including: The multi-dimensional raw power data collected by the smart meter are acquired, including: Current data, voltage data, power data, power consumption duration data, and electric meter equipment identification information; The data format is time series structured data: , wherein denotes the data collection time step, each data contains a power parameter dimension, i.e. ; The original power data is cleaned to eliminate invalid data caused by missing values, abnormal isolated points and data collection errors. The missing value elimination adopts a neighborhood interpolation verification method. If the same parameter data of a continuous time step is missing, the time sequence of this segment is directly eliminated. The isolated point elimination adopts a quartile range method to calculate the quartiles , and the quartile range of the parameter data. When the data value is out of the range , it is determined as an isolated point and is eliminated. The cleaned power data is standardized. The standardization expression is: , wherein denotes the original data of the th power parameter at the th time step, is the mean value of the th power parameter, is the standard deviation of the th power parameter, is the cleaned data time step, , .

[0027] In this way, the multi-dimensional original data collected by the smart meter is obtained and cleaned to eliminate missing values, isolated points, and then standardized. The cleaning can remove invalid data, ensure data quality, and avoid interference on subsequent analysis. The standardization processing makes different parameter data in the same dimension, which is convenient for comparison and calculation, improves data usability, and adopts a specific method to process missing values and isolated points, which is scientific and reasonable, can maximize the retention of effective information, and lays a solid foundation for building accurate knowledge graph and precise subsequent operations such as anomaly detection.

[0028] In an embodiment, a power field knowledge graph is constructed based on the preprocessed power data, comprising: Defining the core entity types of the knowledge graph, including smart meter device entities, power parameter entities, fault type entities, device association entities and time entities; Building the association relationship between entities, including the collection relationship between the meter device and the power parameter, the connection relationship between the meter device and the device association entity, the triggering relationship between the power parameter and the fault type, and the corresponding relationship between the fault type and the historical record; The knowledge graph data is stored in a graph database, and the preprocessed power data is mapped as entity attribute values in the knowledge graph, and the correlation weight between entities is calculated: , wherein is the correlation frequency of the entity and the entity , is the total number of entities associated with the entity , to form a complete power field knowledge graph containing nodes, edges, attributes and correlation weights, wherein the nodes represent various entities, the edges represent the correlation between entities, and the attributes include power parameter specific values, equipment models, fault occurrence times and the like.

[0029] In this way, the core entity types are defined, the correlation relationship is constructed, the correlation weight is stored and calculated in the graph database, and the various elements in the power field and their relationships are clearly presented. The scattered power data is structured, the graph database storage is conducive to efficient query and management, the calculation of the correlation weight can reflect the closeness of the relationship between entities, and the complete knowledge graph provides comprehensive and accurate information for anomaly detection, reasoning and the like, so that the system can perform deep analysis based on rich knowledge, and improve the accuracy of fault judgment and processing.

[0030] In an embodiment, the anomaly detection unit fuses machine learning algorithms and knowledge graph features to monitor abnormal fluctuations in power data in real time, including: Extracting entity correlation features, power parameter time series features and historical fault correlation features from the knowledge graph, and calculating the trend value of the time series features: , wherein is a preset time window step, and a multi-dimensional feature set is formed in combination with the entity correlation weight , is the historical fault correlation feature; An anomaly detection model fusing machine learning algorithms is constructed, taking the multi-dimensional feature set as input, and the machine learning algorithms include an ensemble learning model based on XGBoost and an LSTM-AE model for time series anomaly detection. The loss function of the LSTM-AE model is: , wherein is the reconstructed power parameter data of the model, The objective function of the XGBoost model is: , is the loss term, is the regularization term; An anomaly score is calculated by the anomaly detection model: , wherein, is a weight coefficient, is the reconstruction loss of the LSTM-AE model at time t, is the abnormal probability output by the XGBoost model at time t; The determination rule comprises the following steps: Two key evaluation indexes, a parameter time series stability coefficient and an entity correlation strength coefficient , are extracted from the knowledge graph. wherein, The calculation is performed through the formula: , M is the number of parameters, T is the length of the time series, and the smaller the value, the more stable the time series is; The average weight value of the correlation edges between entities in the knowledge graph is adopted, and the larger the value, the closer the correlation is; The final value of is obtained through normalization processing: , wherein , is the extreme value of the time series stability coefficient, , is the extreme value of the correlation strength coefficient, is a domain experience weight, taking a value of 0.6, which is verified to be optimal through 1000 groups of power historical data; Specific adaptation scenarios are as follows: when the time series stability coefficient of the voltage parameter is much lower than the average value 0.2, it is automatically adjusted to 0.7 to strengthen the time series detection advantage of the LSTM-AE; when the correlation strength coefficient of a certain regional electric meter and transformer (higher than the average value 0.5), it is reduced to 0.3 to highlight the mining ability of XGBoost on correlation features; when , is a preset abnormal score threshold, and it is determined that there is abnormal fluctuation of power data; Abnormal fluctuations include current mutation, voltage sudden rise and sudden drop, abnormal deviation of power, and abnormal fluctuation of power load, the model outputs an abnormal detection result and an abnormal confidence: , is the historical maximum abnormal score; The training process of the anomaly detection model combines historical fault data in the knowledge graph as labeled samples, and realizes model accuracy improvement through iterative optimization, surpassing the traditional anomaly detection method based on fixed threshold.

[0031] In this way, multiple features are extracted to construct a multi-dimensional feature set, and two machine learning models are fused to calculate anomaly scores, which combines the advantages of different models. XGBoost is good at processing structured data, and LSTM-AE is suitable for time series data, which can fully capture abnormal features. Combined with the knowledge graph features, the correlation between power data is considered, and the anomaly is accurately determined by calculating the anomaly score and confidence, which surpasses the traditional fixed threshold method, improves the accuracy and flexibility of anomaly detection, and timely discovers abnormal fluctuations in power data.

[0032] In an embodiment, the graph reasoning unit is used to infer the possible fault causes corresponding to the abnormality based on the correlation and reasoning mechanism of the knowledge graph combined with historical data, including: A graph reasoning rule library is constructed, including mapping rules of power parameter abnormalities and fault types, device correlation and fault conduction rules, and historical fault recurrence correlation rules; Based on the entity correlation path in the knowledge graph, a mechanism combining path reasoning and semantic reasoning is used to calculate the path reasoning confidence: , wherein is the edge set in the correlation path, is the correlation weight corresponding to the edge, and the associated entities and potential influencing factors of the abnormal power data are traced back; Fusion of historical fault handling data and real-time abnormal features, calculation of posterior probability of each possible fault cause by Bayesian probability reasoning algorithm:

[0033] wherein, is the prior probability of fault , calculated by historical fault frequency, is the conditional probability of abnormal when fault occurs, and the Top-N fault cause list is output after sorting the posterior probability, wherein N is the preset number of fault cause display.

[0034] In this way, the reasoning rule base is constructed, path and semantic reasoning are combined, and the posterior probability is calculated by using the Bayesian algorithm. The rule base constructed in this way can systematically cover the power failure reasoning rules. The combination of path and semantic reasoning can comprehensively trace abnormal associated entities and potential factors. The Bayesian algorithm can fuse historical and real-time data, scientifically calculate the posterior probability of each failure cause, and output the result after sorting, thereby providing a reliable basis for failure diagnosis, helping operation and maintenance personnel to quickly locate the root cause of failure, and improving the efficiency of failure handling.

[0035] In an embodiment, the diagnostic report generation unit automatically generates a diagnostic report containing abnormal details, failure causes and treatment suggestions, including: Extracting key information in the abnormal detection result, including abnormal occurrence time, abnormal power parameter type, abnormal value range and abnormal confidence ; The failure cause list output by the correlation graph reasoning unit is matched with the failure treatment scheme library stored in the knowledge graph to match the standard treatment suggestions corresponding to each failure cause, and the matching degree of the treatment suggestions is calculated:

[0036] Among them is the semantic similarity between the failure and the treatment scheme , is the maximum semantic similarity; According to the preset report template, the abnormal details, failure cause analysis, posterior probability sorting and treatment suggestions are integrated to automatically generate a structured diagnostic report, which supports text format export and system review.

[0037] In this way, the key information of the abnormality is extracted, the failure causes are matched with the treatment suggestions, and the report is generated according to the template. In this way, the abnormal detection and reasoning results can be automatically integrated, a structured report can be quickly generated, the key information can be extracted to facilitate the staff to quickly understand the abnormal situation, the failure causes can be matched with the treatment suggestions to provide targeted solutions, the report supports export and review, and information sharing and archiving are facilitated, thereby providing comprehensive and standardized reference for power failure handling and improving the standardization and efficiency of operation and maintenance work.

[0038] In an embodiment, the intelligent alarm unit sends corresponding intelligent alarm signals according to the abnormal level, including: Based on the abnormal confidence and the severity coefficient of the failure cause , The abnormal level score is calculated according to the power operation and maintenance specification:

[0039] According to the value of , the abnormal level is divided: When is a slight anomaly, When is a general anomaly, When is a serious anomaly; The corresponding alarm mode is configured for different anomaly levels, including system pop-up prompt, SMS notification, email reminder, and sound and light alarm, wherein the serious anomaly triggers multi-channel synchronous alarm; The alarm signal contains a diagnostic report link and an anomaly core information summary, which supports the staff to quickly locate the anomaly and view the detailed diagnostic content, and realizes the rapid response and processing of power failure.

[0040] In this way, the level score is calculated based on the anomaly confidence and severity coefficient, the level is divided and the alarm mode is configured, the level is scientifically divided according to the actual situation of the anomaly, different levels correspond to different alarm modes, the emergency degree of serious anomaly is highlighted, multi-channel synchronous alarm ensures that the staff receives information in time, the alarm signal contains a diagnostic report link and a core summary, which facilitates the staff to quickly locate and view the detailed content, realizes the rapid response and processing of power failure, and guarantees the stable operation of the power system.

[0041] Those skilled in the art can understand that all or part of the steps of the above-mentioned embodiment methods can be completed by programs instructing related hardware, therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program codes.

[0042] The above embodiments are described in detail, and the principles and implementation manners of the present application are described by applying specific examples; the above embodiment descriptions are only used to help understand the method of the present application and its core idea; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range will be changed; in summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A knowledge graph construction and analysis system for smart meter data, characterized in that, Includes the following units: The system includes a knowledge graph construction unit, an anomaly detection unit, a graph reasoning unit, a diagnostic report generation unit, and an intelligent alarm unit. The knowledge graph construction unit acquires and preprocesses multi-dimensional power data from smart meters, and then constructs a power domain knowledge graph based on the preprocessed data, which includes information on related meter devices, power parameter data, fault records, and device relationships. The anomaly detection unit integrates machine learning algorithms and knowledge graph features to monitor abnormal fluctuations in power data. The graph reasoning unit combines knowledge graph relationships, reasoning mechanisms, and historical data to infer possible causes of failures corresponding to anomalies. The diagnostic report generation unit automatically generates a diagnostic report containing details of the anomaly, the cause of the fault, and handling suggestions; The intelligent alarm unit issues corresponding intelligent alarm signals based on the level of abnormality, thereby realizing power anomaly detection and fault management.

2. The knowledge graph construction and analysis system for smart meter data according to claim 1, characterized in that, The multi-dimensional power data acquired by the knowledge graph construction unit includes: The data must be at least one of the following: current data, voltage data, power data, electricity usage duration data, and meter identification information, and the data type must be time-series structured data.

3. The knowledge graph construction and analysis system for smart meter data according to claim 1, characterized in that, The knowledge graph construction unit preprocesses multi-dimensional power data, including: Data cleaning and standardization; Data cleaning is used to remove missing values, outliers, and invalid data caused by data acquisition errors. Standardization is used to unify the data dimensions of different power parameters.

4. The knowledge graph construction and analysis system for smart meter data according to claim 1, characterized in that: The knowledge graph in the power sector includes core entity types and relationships between entities; The core entity types include at least one of the following: smart meter device entity, power parameter entity, fault type entity, device associated entity, and time entity; The relationships between entities include at least one of the following: collection relationship, connection relationship, trigger relationship, and correspondence relationship.

5. The knowledge graph construction and analysis system for smart meter data according to claim 1, characterized in that: The knowledge graph construction unit uses a graph database to store knowledge graph data; The preprocessed power data is mapped to the attribute values ​​of the corresponding entities in the knowledge graph; Simultaneously, the association weights between entities are calculated to improve the knowledge graph.

6. The knowledge graph construction and analysis system for smart meter data according to claim 1, characterized in that, The machine learning algorithm fused by the anomaly detection unit includes the following steps: Models based on ensemble learning and time-series anomaly detection models; The anomaly detection unit extracts entity association features, power parameter time-series features, and historical fault association features from the knowledge graph to form a multi-dimensional feature set; It is then fed into an anomaly detection model built using a machine learning algorithm; Calculate the anomaly score and compare it with a preset threshold; Determine if there are any abnormal fluctuations in power data.

7. The knowledge graph construction and analysis system for smart meter data according to claim 1, characterized in that, When the graph inference unit infers the possible causes of the fault corresponding to the anomaly, it includes the following steps: Construct a reasoning rule base that includes mapping rules between abnormal power parameters and fault types, rules for equipment association and fault propagation, and rules for association of historical fault recurrence; Employ at least one of the following reasoning mechanisms: path reasoning, semantic reasoning, and probabilistic reasoning; By combining historical fault handling data with real-time anomaly characteristics, the causes of faults can be inferred.

8. The knowledge graph construction and analysis system for smart meter data according to claim 1, characterized in that, The diagnostic report generation unit generates a diagnostic report, including the following steps: Extract key information from the anomaly detection results, including the time of anomaly occurrence, the type of abnormal power parameter, the range of abnormal values, and the confidence level of the anomaly. A list of fault causes output by the association graph inference unit; Match the corresponding processing suggestions in the fault handling solution library of the knowledge graph; A structured diagnostic report is generated by integrating information according to a preset report template.

9. The knowledge graph construction and analysis system for smart meter data according to claim 1, characterized in that, The intelligent alarm unit determines the anomaly level by including the following steps: Anomaly level score is calculated based on the anomaly confidence level and the severity coefficient corresponding to the cause of the fault in the anomaly detection results. Based on the score, the abnormality is classified into at least one of the following categories: minor abnormality, general abnormality, and severe abnormality; Configure corresponding alarm methods for different anomaly levels.

10. The knowledge graph construction and analysis system for smart meter data according to claim 1, characterized in that, The alarm methods include: At least one of the following: system pop-up prompts, SMS notifications, email alerts, and audible and visual alarms; Severe anomalies trigger simultaneous alarms across multiple channels. The alarm signals include access to the diagnostic report and a summary of the core anomaly information, enabling quick location of the anomaly and viewing of diagnostic details.

Images